When backing up data it may be desirable to exclude certain files from being backed up. For example, if the backup is being performed to protect user data, such as files or other objects created by the user using one or more applications, it may be desirable to exclude operating system files from being backed up, to avoid using associated storage space for example. Or, in backing up virtual machine data it may be desirable to back up operating system and/or virtual machine state-related files but not application data, such as Exchange Server objects, which may already be backed up by another system and/or process.
In a file-based backup, typically a list of files to be excluded from the backup is created and used to skip the files on the exclude list as the backup application walks the file system to create the backup. In a block-based backup, by contrast, the backup is performed at the volume level and it is not as simple a matter to skip files on the exclude list. Instead, it is necessary to determine which blocks store file data and/or metadata associated with files to be excluded, and to omit those from the backup data. On incremental backup, moreover, it is necessary to determine which blocks modified on the source volume are associate with files (data and/or metadata) that have been excluded from the prior backup(s) and for consistency are desired to continue to be excluded from the current incremental backup.